Article below is written with support of ChatGPT, and lightly edited by me
It has been frustrating, to say the least, to see my portfolio significantly underperform against the S&P 500 this year. As shown above, the gap between my portfolio’s performance and the S&P 500 is the largest it has ever been since I began directly investing in stocks in 2018. This underperformance has also resulted in a steep contraction in my portfolio’s value.
However, as David Teppler of Appaloosa Investment Management pointed out in a recent interview, it is important for retail investors to “look past” these short-term setbacks and focus on the long-term value of their investments. With this in mind, my observations on the market are set out below. I will provide more updates on the performance of the portfolio’s companies in Q1 2023 when their results come in.
The decline in value of the overall stock market in 2022 has been particularly punishing for some stocks, especially the so-called “tech” stocks. Regardless of the underlying quality of the business, the labeled “tech” stocks that I own, notably Amazon, Google, and Meta, have seen their multiples on a normalized earning basis decline from around 25x to the low to mid-teens multiple and all the way to a single-digit multiple for Meta.
This naturally raises the question of whether 25x was excessive for these companies?
One could justify a 25x multiple if the following conditions are met:
a/ That It is likely, as far as the eyes can see, that these businesses will exist for a long time (long enough that the terminal value comprises a small portion of today’s present value);
b/ The discount rate and the long-term growth rate adopted by the investor justify it. In my case, I used a 6% discount rate and a 2% long-term growth rate, and hence perpetuity math implies a 25x multiple.
One could challenge point (b) and argue that a 6% discount rate in a world where the T-bill is closing in on 4% presents too small a margin of safety, which I would somewhat agree to. If I put in a 100% margin of safety and revise my discount rate upwards to 8% while retaining a 2% long-term growth rate, the implied multiple is now 16x, i.e. a contraction in value of 36.0%, ceteris paribus.
So, in light of the rapid increase in the base rate, which is now at 4.25% and potentially rising to 5.25% as expected by the market, it is understandable why the multiple has contracted. While these rates are technically still below my original assumption of 6%, they do present a much smaller margin of safety.
For prudence, I have adopted an 8% discount rate while retaining a 2% long-term growth rate. In this environment, an attractive multiple to buy stocks at would therefore be around 16x or less.
I will discuss the companies in the portfolio in the next quarter, in anticipation of their Q1 2023 results.
Reading / listening recommendation – Artificial Intelligence
This quarter, much of my reading and listening has been about artificial intelligence (AI). In a personal capacity, other tech presents two challenges for me to grasp: (a) they could be fads (e.g. cryptocurrency, note NOT the underlying blockchain technology) or (b) they are real but way too difficult for me to understand even if I spend a great amount of time on them (e.g. biotech, nuclear fusion, etc).
However, AI is real and it is possible that, with a good amount of time spent, I can understand, in a general way, where the field is heading.
The following is a list of reading and listening materials that are shaping my way of thinking on AI:
- Interviews with Yann Lecunn
- Interviews with Andrew Ng
- Ben Thompson’s articles on everything related to AI
- Meta AI research lab
- Section on AI and machine learning on Amazon’s science website
- Eye on A.I. podcast by Craig S. Smith of the New York Times
- Developments such as generative AI (text, images, videos), within text large language models, and computer vision.
With cautious optimism, I am very excited to see that Amazon, Google, and Meta – all of which I own stocks in my portfolio – are not only doing leading research work but also churning out a lot of AI-as-a-service capabilities for external clients (for the first two) and applying AI and machine learning in augmenting existing services (for Meta, including work such as a video recommendation engine for Reels and applied AI/ML in targeting ads).